JP2812203B2 - Image playback device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coded image transmission system and an image reproducing apparatus, and more particularly to a transmission system for transmitting an image code compressed and coded by using a discrete cosine transform (DCT) as an orthogonal transform and its code. The present invention relates to an image reproducing apparatus for reproducing a coded image.

[0002]

2. Description of the Related Art A digital image signal is converted to a CD-RO.
When recording on a recording medium such as a hard disk, a hard disk, or a magnetic tape, the data amount is enormous, and therefore, it is usually recorded by compression encoding.

[0003] Various compression coding methods have been proposed. Transform coding using an orthogonal transform that efficiently utilizes the correlation in a two-dimensional space of an image is often used. In particular, a DCT based orthogonal transform is used. The encoding method used was the color still image standard JPEG (Joint Photograph).
hic Expert Group) and moving picture coding standards for storage media (MPEG: Moving Pict)
(Us Expert Group).

[0004] Reproduction of an image code by a conventional DCT-based coding method will be described using MPEG as an example. FIG. 13 is a block diagram showing an example of a conventional image reproducing apparatus. This conventional image reproducing apparatus reads an image code conforming to the MPEG and analyzes the type of the code by the header analysis means 131.

In MPEG, one sequence of a moving image is divided into a group of pictures (GOP) consisting of a plurality of frames (pictures) and encoded, as will be described in detail later. The GOP is an I-picture that is an intra-coded image, a P-picture that is an inter-coded image predicted from a temporally previous frame that has already been encoded, and an inter-frame that is predicted from two temporally preceding and succeeding frames. It is composed of a total of three types of B pictures, which are coded images.

When the input image code is analyzed as an I-picture by the header analyzing means 131, the variable-length Huffman code obtained by compressing the I-picture with high efficiency is decoded.
2, is inversely quantized by the inverse quantization means 133, and further inversely DCT (I
The value of the image of the block is calculated by performing DCT) processing, the image is decompressed, output via the adder 135, and input to the previous frame unit 136 or the rear frame unit 137 and stored.

When the input image code is analyzed as a P picture by the header analyzing means 131, the P picture is decoded by the decoding means 132, inversely quantized by the inverse quantization means 133, and further transmitted by the IDCT means 134.
The difference value of the block is calculated by the CT process and the adder 1 is calculated.
35, where it is added to the image code of the previous frame stored in the previous frame unit 136 by the previous prediction unit 138 and the motion-compensated block, expanded to the original image and output. The data is input to and stored in the frame unit 136 or the rear frame unit 137.

When the input image code is analyzed as a B picture by the header analyzing means 131, the B picture is decoded by the decoding means 132, inversely quantized by the inverse quantization means 133, and further transmitted by the IDCT means 134.
The difference value of the block is calculated by the CT process and the adder 1 is calculated.
35, where the two prediction means 139 apply motion compensation to the image codes of the preceding and succeeding frames stored in the previous frame part 136 or the subsequent frame part 137, and the rear frame part 137 The stored image code of the subsequent frame is added to the block that has been subjected to the motion compensation, expanded to the original image and output, and input to the previous frame unit 136 or the rear frame unit 137 and stored.

As described above, any playback device based on the international standard MPEG can be used as the MPEG playback device.
Can be reproduced. But,
In the processing of the decoding means 132 and the IDCT means 134, the load on the central processing unit (CPU) to be used is large, so that high-speed reproduction cannot be performed unless a high-speed CPU is used.

For example, in order to perform a processing of 15 frames / second in an image code reproducing apparatus conforming to JPEG or MPEG,
It is necessary to perform the reproduction process of one frame in about 66 ms.
If it takes 30 ms to decode the Huffman code, 10 ms to dequantize, 20 ms to the IDCT processing by the IDCT means 134, and 20 ms to display, the entire processing time is about 80 ms. It takes 14 ms to reproduce the frame image.

Therefore, in order to reduce the load on the CPU,
Conventionally, a method that allows a high-speed reproduction even with a low-speed CPU by changing a part of the encoding method to calculate the difference between neighboring pixels instead of the IDCT calculation or by omitting quantization. More various proposals have been made.

For example, in the reproducing apparatus described in Japanese Patent Application Laid-Open No. 2-241270, a pixel not to be coded is provided between pixel blocks, and pixels between blocks are supplemented by referring to adjacent pixel blocks during reproduction. Reproduction can be performed without burdening the CPU.

In the reproducing apparatus described in Japanese Patent Application Laid-Open No. 4-971, a transform coefficient of a pixel block is divided into a plurality of bands based on a spatial frequency and subjected to variable-length coding for each region.
The code data obtained by the variable length coding is stored for each region, and the amount of code data is managed for each region. The number of regions is determined so that the total code data amount does not exceed a predetermined code data amount. Perform variable-length decoding of the transform coefficient data of each area during reproduction using only the code data of each area of the number of areas, and perform two-dimensional inverse orthogonal transform with the transform coefficients of the high-frequency areas more than the determined number of areas being zero. As a result, the processing amount is reduced.

In order to reduce the load on the CPU,
A method of deleting information of a portion where deterioration of an image is less noticeable when encoding is also conceivable. For example, JP-A-3-25
No. 6452 discloses that a digital image is divided into a plurality of blocks, transform coefficients obtained by performing DCT processing on each of the blocks are divided by respective threshold values of a quantization matrix, and quantization is performed. In a data compression method in which the coefficients are converted into a one-dimensional sequence from the DC component to the high-frequency component in a fixed order and the number of consecutive zeros in the sequence obtained by the conversion is encoded, the DCT process or the quantum By setting a control variable for each block for the transformed conversion coefficient, and performing a kind of low-pass filter processing that makes components parallel to the diagonal line indicated by the control variable and higher high-frequency components to zero, For the peripheral part of the image that is not so important, the value of the control variable is reduced to reduce high frequency components, so that the amount of data after compression is reduced. Over data processing method is disclosed.

Japanese Patent Application Laid-Open No. 4-357788 discloses that a frame of a moving image is divided into a predetermined number of blocks, each block is divided into low-frequency components and other components, and hierarchical coding is performed. In a moving image encoding apparatus that performs encoding using inter-frame prediction or motion compensation, an image block to be transmitted is compared with an image decoded from only a layer corresponding to a low-frequency component of a layered code of a previous frame. The inter-frame prediction or the motion compensation is performed, and the motion vector is zero as a result of the motion compensation, and the error obtained by comparing the transmission block with the block of the image decoded using the layered code of the previous frame is obtained. There is disclosed an apparatus in which the transmission block is not transmitted when the threshold is equal to or less than a predetermined threshold, thereby eliminating code redundancy, that is, reducing the code amount.

Japanese Patent Application Laid-Open No. 4-299688 discloses that a coded signal obtained by coding a difference between a reproduced image one frame before and a predicted value is variable-length coded, and the motion detection unit performs In a high-efficiency video coding method in which a motion vector detected by performing matching between a reproduced image and an original image is subjected to variable-length coding using a motion compensation coding table, a motion vector among a plurality of motion compensation coding tables is used. An encoding method has been disclosed in which a motion compensation encoding table is selected in accordance with the nature of the frequency of occurrence of a vector, thereby reducing the mismatch of a motion vector and performing efficient encoding.

[0017]

However, since all of the above-mentioned conventional image reproducing apparatuses and encoding systems perform image compression for changing a part of the encoding system, an international standard encoding system is used. However, there is a problem that a dedicated image reproducing device is required because the compatibility with the image cannot be obtained.

In the conventional image compression for reducing the amount of image compression codes, it is necessary to determine an optimal compression method for each block, so that it takes time to generate compression codes. Further, since the code amount does not correspond to the processing capacity of the image reproducing apparatus, real-time reproduction cannot be performed by a reproducing apparatus having a low processing speed. Furthermore, since the compression code amount is fixed, there is a problem that even a high-speed reproduction device can reproduce only a code having a predetermined image quality.

Therefore, the present invention provides a DCT with a simple configuration without impairing the image quality as much as possible in accordance with the capability of the reproducing apparatus.
It is an object of the present invention to provide a coded image transmission system and an image reproducing apparatus capable of reproducing a code based on the coded image at a high speed.

[0020]

In order to achieve the above object, according to the first aspect of the present invention, there is provided an encoded image transmission system which divides an image into a plurality of blocks and uses a discrete cosine transform for each block. Intra-frame coded image compressed and coded, and for each block, search for a block in which the difference is the smallest in the current frame and both before and after the current frame, and perform motion compensation. The inter-frame coded image obtained by compression-coding the difference value between the current frame block and the motion-compensated block by using the discrete cosine transform is provided at the beginning of the entire coded image sequence.
In a coded image transmission method of multiplexing and transmitting together with a sequence header to be transmitted, at least one of a type of an inter-frame coded image and a value of motion compensation to be transmitted together with an intra-frame coded image is fixed to identify a fixed parameter. This is transmitted together with the sequence header having an identifier to be made, and the reproducing side detects the identifier in the header to identify the fixed parameter of the input image code and perform the decompression process.

According to the image reproducing apparatus of the present invention, an intra-frame coded image obtained by dividing an image into a plurality of blocks and compression-encoding each block using a discrete cosine transform, A motion compensation is performed by searching for a block having the smallest difference between the current frame and a frame temporally earlier than the current frame for each block, and performing motion compensation on the difference between the current frame block and the motion compensated block. The inter-frame coded image compressed and coded using the discrete cosine transform uses a fixed parameter identifier for all coded image sequences.
User in the sequence header provided at the beginning of the sequence.
An image reproducing apparatus for decompressing and reproducing an image code multiplexed with the sequence header included in the data area, wherein the input image code and the inter-frame code are based on a sequence header in the input image code. And a header analysis unit that determines whether the type of the inter-frame coded image is a code having a fixed type. The input image code is a code in which the type of the inter-frame coded image is fixed by the header analysis unit. When it is determined that there is an input image code, it is detected whether or not the input image code is an intra-frame encoded image. When the input image code is detected as an intra-frame encoded image, expansion processing of the intra-frame encoded image is performed. And a decompression processing means for decompressing the inter-frame coded image when it is detected.

In the image reproducing apparatus according to the third aspect,
The input image code according to claim 2, wherein a transmission interval between the intra-frame encoded image and the next intra-frame encoded image is constant, and the inter-frame encoded image is transmitted within the transmission interval. Things.

Further, in the image reproducing apparatus according to the fourth aspect , at the beginning of the entire coded image sequence in the input image code ,
Based on the provided sequence header , a header analysis unit that determines whether the input image code is a code having an inter-frame coded image with fixed motion compensation, and the input image code is fixed with motion compensation by the header analysis unit. When it is determined that the input image code is a decoded code, the image processing apparatus includes a decompression processing unit that performs an input image code decompression process by omitting the motion compensation process.

Further, in the image reproducing apparatus according to the fifth aspect,
Decompression processing means for decompressing and reproducing an input image code; difference value total calculation means for calculating a cumulative value for each frame of a difference value of the inter-frame coded image decompressed by the decompression processing means; And control means for stopping the updating of the display of the output image data of the decompression processing means on the display unit until the cumulative value exceeds the threshold value.

[0025]

According to the coded image transmission method of the present invention, one or both parameters of the type of the inter-frame coded image and the value of the motion compensation are fixed , and a sequence having an identifier for identifying the fixed parameter in the user data area. The header is transmitted together with the header , and set on the playback side at the beginning of the entire encoded image sequence.
Since the decompression process is performed by detecting the fixed parameter of the input image code by detecting the identifier in the sequence header to be input, the type of the inter-frame coded image in the input image code can be specified. Processing for unnecessary inter-coded images that are not performed or motion compensation processing can be eliminated.

Further, in the image reproducing apparatus according to the second aspect,
Since the type of the inter-frame coded image input together with the intra-frame coded image is only the forward inter-frame coded image, if the type is not the intra-frame coded image, it is a forward inter-frame coded image. The process of determining the type of the inter-frame coded image and the process of decompressing the inter-frame coded image in both the forward and backward directions can be eliminated.

Further, in the image reproducing apparatus according to the third aspect,
Since the transmission interval of the intra-frame coded image is constant, the type of the coded image can be determined based on the frame interval without using the header.

Further, in the image reproducing apparatus according to the fourth aspect,
It is determined whether or not the input image code is a code having an inter-frame coded image with fixed motion compensation.If the input image code is a code with an inter-frame coded image with fixed motion compensation, the motion compensation process is omitted. Since the input image code is decompressed in this way, processing time for motion compensation can be eliminated.

Further, in the image reproducing apparatus according to the fifth aspect,
The threshold value is compared with the accumulated value of the sum of the difference values of the inter-frame coded images decompressed by the decompression processing means for each frame, and the output image data of the decompression processing means is output until the accumulated value exceeds the threshold value. Since the updating of the display on the display unit is stopped, the image of the previous frame is displayed as it is until the accumulated value exceeds the threshold, and the time for displaying the inter-frame coded image at that time is reduced. Can be omitted.

[0030]

Next, an embodiment of the present invention will be described. FIG. 1 is a block diagram showing an embodiment of an image reproducing apparatus according to the present invention. The image reproducing apparatus according to the present embodiment includes a header analysis unit 11 for analyzing an input image code, a decoding unit 12 for decoding the input image code, an inverse quantization unit 13 for performing inverse quantization of the decoded image data, Inverse DC of quantized image data
IDCT means 14 for performing T processing, an adder 15, a previous frame section 16 in which image data one frame time before the output of the adder 15 is stored, and an output of the adder 15 A rear frame unit 17 in which image data one frame later in time is stored, a pre-prediction unit 18 for performing prediction processing based on an output of the previous frame unit 16, and both outputs of the front frame unit 16 or the rear frame unit 17 Prediction means 19 for performing a prediction process based on the data, rear prediction means 20 for performing a prediction process based on the output of the rear frame unit 17, and pre-prediction means 2
1 and an adder 22.

In the reproducing apparatus of this embodiment shown in FIG.
The input image is reproduced according to G. So first, MP
The hierarchical structure of the EG code format will be described with reference to FIG. In MPEG, the top layer is a moving image sequence (video sequence), and is composed of a plurality of group of pictures (GOP) as shown in FIG. One GOP is composed of a plurality of frames (pictures) as shown in FIG. The GOP is an I-picture that is an intra-coded image, a P-picture that is an inter-coded image predicted from a temporally previous frame that has already been encoded, and an inter-frame that is predicted from two temporally preceding and succeeding frames. It is composed of a total of three types of B pictures, which are coded images.

Each picture is composed of a plurality of slices divided into arbitrary regions as shown in FIG.
The slice is composed of a plurality of macroblocks arranged in order from left to right or from top to bottom as shown in FIG. As shown in FIG. 2E, one macro block includes four adjacent luminance component blocks Y1 to Y.
4 and one chrominance component block Cb and one chrominance component block Cr in a region corresponding to the position.

Further, each of these blocks is composed of 8 × 8 pixels as shown in FIG. This block is the minimum unit of coding, and is the processing unit of DCT. Note that the minimum unit of motion prediction compensation is a macroblock.

Next, the format of an input image code conforming to MPEG will be described with reference to FIG. The input image has a sequence header Seq for each layer, a GOP header Gop added for each GOP, a picture header for identifying each picture, a slice header for identifying each slice, and a slice header for identifying each macroblock. , A macro header Macro and a block code Block.

Here, the head of the input image code, that is,
As shown in FIG. 3, the sequence header Seq arranged only at the beginning of the first frame includes a start code SeqStart of the sequence header, a horizontal number indicating the number of pixels (horizontal size) after expansion in the horizontal direction, and a horizon after expansion in the vertical direction. Indicating the number of pixels (vertical size), a code UsrStart indicating the presence of user data, and user data UserData.

That is, in the MPEG, since the user data area UserData that can be freely defined by the user is included in the sequence header Seq, this embodiment shows that the code is a code compressed with fixed parameters as shown in FIG. For example, a 6-byte identifier "Custom"
For example, there is one feature in that a 2-byte fixed parameter flag “cflg” is arranged.

The above fixed parameter flag cflg is
When the least significant bit (B0) of the two bytes is "1",
Indicates that the image code is limited to I-pictures and P-pictures, and indicates the second least significant bit (B
When 1) is "1", it indicates that the image code has a fixed motion compensation value.

Returning to FIG. 1, the description will be continued. The present embodiment has a configuration capable of performing high-speed reproduction and normal reproduction. At the time of normal reproduction, the apparatus according to the present embodiment uses the intra-frame coded image I
A picture, a P picture that is an inter-frame coded image predicted from a temporally previous frame that has already been coded, and a B picture that is an inter-frame coded image predicted from two temporally preceding and succeeding frames. The type of encoded image (image code) is expanded and reproduced.

When the header analysis means 11 determines the picture type of the input image code, it also outputs a control signal for controlling the addition operation in the adder 15 (not shown in FIG. 1), Both prediction means 19, post prediction means 20
And the output operation of the IDCT means 14 is selectively controlled. This adder 1
The selection control of the addition operation in 5 may be performed by separately outputting a control signal from the IDCT means 14.

When the input image code is an I picture, the I picture is decoded by the decoding means 12 and inversely quantized by the inverse quantization means 13, and then subjected to inverse DCT (IDCT) processing by the IDCT means 14. Then, the value of the image of the block is calculated, the image is decompressed, and further output as it is via the adder 15, and is input to the front frame unit 16 or the rear frame unit 17 and stored.

When the input image code is analyzed as a P picture by the header analysis means 11, the P picture is decoded by the decoding means 12, dequantized by the dequantization means 13, and subjected to IDCT processing by the IDCT means 14. , The difference value of the block is calculated and supplied to the adder 15, where the difference value is added to the image code of the previous frame stored in the previous frame unit 16 by the previous prediction unit 18 with the motion-compensated block, and the original value is added. The image is output after being decompressed into an image, and input to and stored in the front frame unit 16 or the rear frame unit 17.

When the input image code is analyzed as a B picture by the header analysis means 11, the B picture is decoded by the decoding means 12, dequantized by the dequantization means 13, and further IDCT by the IDCT means 14. The difference value of the block is calculated by the processing and supplied to the adder 15, where the prediction unit 19 applies the motion compensation to the image codes of the previous and next frames stored in the previous frame unit 16 or the rear frame unit 17. Or pre-prediction means 18
A block obtained by performing motion compensation on the image code of the previous frame stored in the previous frame unit 16 and a block obtained by performing motion compensation on the image code of the subsequent frame stored in the rear frame unit 17 by the rear prediction unit 20 The image is added and decompressed to the original image and output, and is also input to the front frame unit 16 or the rear frame unit 17 and stored.

The normal reproduction operation described above is the same as the conventional operation, but this embodiment has a different feature from the conventional one in the high-speed reproduction operation described below. In this high-speed reproduction, the header analysis means 11 which reads the input image code and analyzes the header is composed of the user data UserD in the sequence header Seq.
It is detected whether or not “ata” has the identifier “Custom”, and high-speed reproduction is performed when the identifier “Custom” exists.

In the case of high-speed reproduction, processing is performed by assuming that I and P pictures and the value of motion compensation are fixed. In the case of an I picture, the decoding means 12 decodes the picture,
, And the IDCT means 14 calculates the pixel values of the block by IDCT processing to expand the image.

In the case of a P picture, the decoding means 1
2 and inversely quantized by the inverse quantization means 13,
The T means 14 calculates a difference value of the block by the IDCT process, and adds the difference value and the block which has been subjected to the motion compensation to the image code of the previous frame stored in the previous frame unit 16 by the previous prediction means 18, respectively. By adding at 15, the image is expanded and reproduced.

Next, the operation at the time of the high-speed reproduction will be described in more detail. First, a process when the picture configuration is fixed will be described. As described above, there are three types of pictures defined by MPEG: I pictures, P pictures, and B pictures. When reproducing the input image code, it is not necessary to refer to the images of the preceding and succeeding frames because the I picture is an intra-frame coded image. In contrast,
The P picture needs to perform pre-prediction with reference to the picture of the previous picture, and the B picture needs to perform post-prediction or both predictions with reference to the picture of the previous frame or the picture of the subsequent frame.
Therefore, the processing time of the code of the B picture is longer than that of the code of the I picture or the P picture.

Therefore, if the input image code is, for example, as shown in FIG.
As shown in (1), when the first to fourth frames are composed of an I picture, a B picture, a B picture, and a P picture in that order, the first frame is an I picture, and there is no reference frame. Since the second frame and the third frame are B pictures, respectively,
Post-prediction or both predictions are performed by referring to the I-picture image of the frame and the P-picture image of the fourth frame as the succeeding frame as schematically indicated by 41 to 44. Further, the fourth frame is a P picture, and the previous prediction is performed with reference to the image of the I picture of the first frame which is the previous frame. Therefore, in this case, it takes a long time to process the B picture, which is not suitable for high-speed reproduction.

On the other hand, the input image code is fixed to only the I picture and the P picture. For example, as shown in FIG.
When the frame is an I picture and the second to fourth frames are P pictures, the first frame is an I picture and has no reference frame. Since the second to fourth frames are P-pictures, respectively, as shown schematically in FIG. Make predictions only. Therefore, in this case, since the processing of the B picture is unnecessary, it is suitable for high-speed reproduction.

The operation at the time of reproduction of the embodiment of FIG. 1 when an image code having the structure shown in FIGS.
This will be described with reference to the flowcharts of FIGS.
First, when an image code having the picture configuration shown in FIG. 4A is input, the operation is performed according to the flowchart shown in FIG.

That is, the header analysis means 11 shown in FIG. 1 uses the picture header (“Pictur
e "), the picture type is read (step 5).
1), determine whether it is an I picture (step 5)
2) In the case of an I picture, the above-described processing of the I picture is performed by the encoding means 12, inverse quantization means 13 and IDCT means 14 in FIG.

If the header analyzing means 11 determines in step 52 that the picture is not an I picture, it determines whether or not the picture is a P picture (step 54). If the picture is a P picture, it processes the P picture (step 55). Is determined to be a B picture, and the B picture is processed (step 56). The processing of the P picture is performed by the encoding means 12, the inverse quantization means 13, the IDCT means 14,
The adder 15, the previous frame unit 16, the rear frame unit 17, and the previous prediction unit 18 perform the processing as described above, and the processing of the B picture is further performed including both the prediction unit 19 and the rear prediction unit 20.

On the other hand, when an image code fixed to only the I picture and the P picture is input as shown in FIG. 4B, the operation according to the flowchart of FIG. 5B is performed. In the figure, the same reference numerals are given to the same processing steps as those in FIG. That is, first, the header analysis means 11 of FIG.
"Seq") to the fixed parameter identifier "Cust"
om "is detected, the type of picture is read from the picture header (" Picture "in FIG. 3) (step 57), and it is determined whether or not it is an I picture (step 52). Encoding means 12, inverse quantization means 13 and IDCT means 14 of FIG.
To perform the above-described I-picture processing and perform extended reproduction (step 53).

When the header analyzing means 11 determines in step 52 that the picture is not an I picture, the picture is a P picture since the input image code is a fixed parameter picture code, and the P picture is Processing is performed (step 58).

As described above, according to the present embodiment, by fixing the picture structure to the I picture and the P picture,
The process of determining whether a picture is a B picture can be omitted, and post-prediction or both predictions of the B picture can be omitted, so that high-speed reproduction can be performed. In this case, since the encoding format itself is the same as MPEG,
Compatible with.

The above is the operation when the type of the picture to be transmitted is fixed to the I picture and the P picture. However, the transmission interval of the I picture may be fixed and the P picture may be transmitted within the transmission interval. . The operation of this embodiment in this case will be described with reference to the flowchart of FIG.

The header analysis means 11 adds "1" to a variable i having an initial value of 0 when detecting a sequence header (step 61), and subsequently determines whether or not the variable i is larger than a key frame rate. (Step 62). The key frame rate is the number of frames indicating a fixed transmission interval of the I picture.

If it is determined in step 62 that i is smaller than the key frame rate, the decoding means 1
2. The difference value obtained by the inverse quantization means 13 and the IDCT means 14 is added to the previous frame 16
Is processed by adding a block obtained by performing motion compensation to the image code of the previous frame stored in the P-picture to expand and reproduce the image (step 6).
5).

Here, since the P picture needs to refer to the picture of the previous picture, the I picture does not need to refer to the picture of the previous picture, the first frame is always the I picture. Since the I-picture interval (key frame rate) is fixed, the flowchart of FIG.
The value of the variable i compared in step 62 represents the number of frames from the I picture to the current frame.

Therefore, if it is determined in step 62 that the value of the variable i has become larger than the key frame rate, it is known that the input image code at that time is an I picture, and the decoding means 12 and the inverse quantization The I picture processing is performed by the means 13 and the IDCT means 14 (step 6).
3). After the processing of the I picture is completed, the above variables are reset to 0 to prepare for the next processing (step 64).

As described above, in the case of the processing of FIG. 6 as well, the processing of determining whether or not the picture is a B picture can be omitted, as in the case of FIG. Since the prediction can be omitted, high-speed reproduction can be performed. In this case, since the picture type is determined using the variable i, the picture type can be determined without reading the picture header.

Next, the processing when the motion compensation is fixed will be described. In MPEG, P-pictures and B-pictures have motion compensation, and the position of a block to be referenced can be moved. That is, in MPEG, an image is divided into small blocks, and the current frame and its 1
A motion compensation is performed by searching for a block having the smallest difference between two frames before or before and after the frame, and taking a difference between the block of the current frame and the motion-compensated block, performing DCT on the difference block, The result of the conversion is quantized to obtain a P-picture or a B-picture, which is a highly efficient inter-frame code.

FIG. 7A shows a case where the motion compensation is not fixed, and FIG. 7B shows a case where the motion compensation is fixed at (0, 0) (no motion compensation).

In the example shown in FIG. 7A, the block P of the previous frame is referred to the motion-compensated block P 'of (dx, dy), added to the difference block D, and added to the block Q of the current frame. Stored. On the other hand, FIG. 7B shows that the block P of the previous frame is referenced, the difference block D is added to the block P, and the block P is stored in the block Q of the current frame.

The operation of the embodiment of FIG. 1 in the case where the motion compensation shown in FIG. 7A is performed will be described with reference to the flowchart of FIG. In the P picture processing with motion compensation, first, the variables y and x are respectively set to initial values 0 (steps 70 and 71), and then the sign of the motion compensation of the block is expanded by the pre-prediction means 18 (step 70). 7
2) The motion compensation value is stored in variables dx and dy (step 73).

Next, the sign of the difference value of the block is expanded by the pre-prediction means 18 (step 74), and the position of the block referred to from the expanded motion compensation is calculated by the following equation (step 75).

X ′ = x + dx (1) y ′ = y + dy (2) Next, the pre-prediction unit 18 extracts the block at the position calculated in step 75 from the previous frame unit 16 (step 76). After adding the difference value to the extracted block (step 77), the block is written to the buffer of the current frame (step 78). Then, the variable x
8 ”is added (step 79), and it is determined whether or not the variable x after the addition is larger than the horizontal size (Horizon in FIG. 3) analyzed from the sequence header Seq (step 80).

Step 7 when x ≦ (horizontal size)
Returning to the process of 2, when x> (horizontal size), the routine proceeds to step 81, where "8" is added to the variable y. Subsequently, it is determined whether or not the variable y after the addition is larger than the vertical size (Vertical in FIG. 3) analyzed from the sequence header Seq (step 82). If y ≦ (vertical size), the process returns to step 71.

When y> (vertical size), the image data of the current frame is output from the adder 15 to a display unit (not shown) and displayed (step 83), and the image data of the current frame from the adder 15 is converted to the previous frame. It is stored in the section 16 (step 84).

Next, the operation of the embodiment of FIG. 1 when there is no motion compensation shown in FIG. 7B will be described with reference to the flowchart of FIG. In the figure, the same reference numerals are given to the same processing steps as those in FIG. When the header analysis unit 11 detects that there is no motion compensation based on the value of the flag cflg in the sequence header Seq, the processing of the flowchart in FIG. 9 is executed.

In the P picture processing without motion compensation, first, the variables y and x are respectively set to initial values of 0 (steps 70 and 71), and the code of the motion compensation of the block is skipped (step 91). The sign of the difference value of the block is expanded by the pre-prediction means 21 (step 9).
2).

Next, the previous prediction means 21
, A block at a position (x, y) represented by variables x and y is extracted (step 93), the expanded difference value is added to the extracted block (step 77), and the block is written to the buffer of the current frame (step 93). Step 7
8). Thereafter, similarly to the processing when there is motion compensation, “8” is added to the variable x (step 79), and the variable x after the addition is added.
Is larger than the horizontal size (step 8).
0).

When x ≦ (horizontal size), step 9
Returning to the processing of 1, when x> (horizontal size), the routine proceeds to step 81, where "8" is added to the variable y. Subsequently, it is determined whether or not the variable y after the addition is larger than the vertical size (step 82). If y ≦ (vertical size), the process returns to step 71.

If y> (vertical size), the image data of the current frame is output from the adder 15 to a display unit (not shown) and displayed (step 83), and the image data of the current frame from the adder 15 is converted to the previous frame. It is stored in the section 16 (step 84).

In FIG. 7B and FIG. 9, the motion compensation is (0, 0), but (2, 2), (-1, 0),
Even with a fixed value such as (0, 5), the position of the reference block in the previous frame is (x + 2, y + 2), (x-1, y), or (x, y + 5), and each reference block is adjacent. It is not necessary to calculate the reference position for each block.

As described above, by fixing the motion compensation, it is possible to omit the process of decompressing the motion compensation code for each block and calculating the position of the block to be referred to, thereby enabling high-speed reproduction.

Next, the process of updating / not updating the display will be described. In MPEG, a code can be composed of an I picture that is an intra-coded image and a P picture that is an inter-coded image. When the difference value is small, even if the P picture image is not displayed, the appearance does not greatly change, so that the display time of the P picture image at that time can be omitted and the speed can be increased.

Whether the difference value is small or not is determined by comparing the difference value with a value (threshold value) determined by the processing speed of the reproducing apparatus. A playback device capable of high-speed processing reduces the threshold value to increase the number of display updates, and a playback device with a low processing speed increases the threshold value to reduce the number of display updates, thereby reducing the processing speed. Can be adjusted.

FIG. 10A is a diagram for explaining the processing when updating the display of all P pictures, and FIG. 10B is a diagram for explaining the processing when the display of P pictures is not updated until a predetermined condition is satisfied. It is. In FIGS. 11A and 11B, “code” is an input image code of the playback device, I and P indicate I and P pictures, respectively, and numbers indicate the order of frames. “Display” indicates a picture displayed by the display unit. The “sum of difference values” below “display” indicates the sum of difference values for each frame.

In the example shown in FIG. 10A, the displayed code matches the input image code, and all the input image codes are displayed regardless of the sum of the difference values. In contrast,
In FIG. 10B, the broken-line rectangle shown in the column of “display” indicates that the display has not been updated. This non-update of the display is realized by keeping the display of the previous frame until the cumulative value of the sum of the difference values for each frame exceeds a predetermined threshold value (here, “54”).

That is, in the example of FIG. 10B, the second frame is a P picture, and the sum of the difference values is the threshold “5”.
Since "28" is smaller than 4 ", the P picture is not displayed and the I picture one frame before is displayed. The next third frame is also a P picture, and the accumulated value of the difference values is" 49 ". (= 28 + 21), the P picture is not displayed, and the I picture of the first frame is continuously displayed.

The next fourth frame is also a P picture, and the sum of the difference values is “15”, the cumulative sum of the difference values is “64” (= 28 + 21 + 15), and the threshold value is “5”.
Since the value is larger than 4 ", the image display of the P picture P4 of the fourth frame is performed (display is updated). By updating this display, the cumulative value of the sum of the difference values is reset to 0. You.

Since the next fifth frame is also a P picture, and the sum of the difference values and the accumulated value are "43" smaller than the threshold value "54", the display of the P picture P5 of the fifth frame is performed. First, the P picture P4 one frame before is displayed. As can be seen from the above description, the P-picture P6 of the next sixth frame is displayed (updated) because the cumulative sum of the difference values exceeds the threshold value "54".

In this embodiment, as described above, the display is not updated until the cumulative value of the difference values of the P-pictures exceeds the threshold value, and the display remains the display of the previous frame. The display is not updated in the third frame, the third frame, and the fifth frame, and the number of display processes for three frames can be reduced, so that high-speed processing can be performed.

Next, the operation of the embodiment of FIG. 1 when updating the display of all the P pictures shown in FIG. 10A will be described with reference to the flowchart of FIG. The header analysis means 11 shown in FIG. 1 reads the picture type from the picture header of the input image code ("Picture" in FIG. 3) (step 111) and determines whether or not it is an I picture (step 112).

In the case of an I picture, the encoding means 1 shown in FIG.
2. The above-mentioned I-picture processing is performed by the inverse quantization means 13 and the IDCT means 14 for decompression reproduction (step 11).
3) Then, it is supplied to a display unit (not shown) via the adder 15 and displayed (step 114). The adder 15
The image data of the I picture expanded from
6 (step 115).

On the other hand, the header analysis means 11
2, if it is determined that the picture is not an I picture, it is determined that the picture is a P picture, and the decoding means 12, the inverse quantization means 13 and the ID
The difference value is expanded by the CT means 14 (step 11).
6) The adder 15 adds the motion-compensated block to the image code of the previous frame stored in the previous frame unit 16 by the previous prediction means 18 (step 117).

As a result, the picture of the P picture is expanded and reproduced, supplied to the display unit for display (step 118), and the expanded image data extracted from the adder 15 is stored in the previous frame unit 16. (Step 11
9).

Next, FIG. 1 shows a case in which the display update of the P picture is not performed until the predetermined condition shown in FIG. 10B is satisfied.
The operation of this embodiment will be described with reference to the flowchart of FIG. 12, the same components as those of the processing steps in FIG. 11 are denoted by the same reference numerals, and description thereof will be omitted.

In FIG. 12, the header analysis means 1 shown in FIG.
1 is a picture header of the input image code ("Pict" in FIG. 3).
ure ”), the type of picture is read (step 1
11), the operation when it is detected as an I picture is shown in FIG.
(Steps 112 to 115), and the I picture is always updated and displayed.

On the other hand, in FIG.
If it is determined at step 112 that the 1 is not an I picture,
Judging that the picture is a P picture, the decoding means 12, inverse quantization means 13, IDCT means 14 and adder 15
The picture image is expanded and reproduced (steps 116, 1
17).

Next, the pre-prediction means 21 in FIG.
Is added to the variable m which is the sum of the difference values of the P pictures from the IDCT means 14 by an internal adder (step 1).
21) It is determined whether or not the value of the variable m, that is, the cumulative value of the sum of the difference values is larger than a predetermined threshold value (step 122).

If the variable m is equal to or smaller than the threshold value, the expanded P-picture image output from the adder 22 is input to the previous frame section 16 (step 125). Therefore,
In this case, since the display memory in the display unit not shown in FIG. 1 is not rewritten by the output image data of the adder 22, the display unit continuously displays the image data of the previous frame stored in the display memory. Becomes

On the other hand, the pre-prediction means 21 in FIG.
When it is determined in step 22 that the variable m is larger than the threshold value, the block obtained by performing motion compensation on the image code of the previous frame stored in the previous frame unit 16 by the previous prediction unit 21 is sent to the adder 22. The P-picture is output and added to the decompressed difference value from the IDCT means 14 to obtain P-picture decompressed image data, which is supplied to the display memory of the display unit to update the contents of the display memory and display the image. Is updated (step 123).

Subsequently, the pre-prediction means 21 resets the variable m to 0 (Step 124). Next, the decompressed image data of the P picture extracted from the adder 22 is stored in the previous frame unit 16 (step 125). in this way,
According to the present embodiment, the display is not updated until the cumulative value of the sum of the difference values between the frames exceeds the threshold value. Therefore, the number of times of displaying an image can be reduced. It is suitable for high-speed playback.

In the above description, only the I picture and the P picture are transmitted.
The process fixed to the picture, the motion compensation fixed process, and the display update / non-update process may be performed independently, or two or all of them may be combined.

Although the above embodiment has been described using MPEG as an example, the present invention is not limited to this, and the present invention is not limited to JPEG or other DCT-based coding schemes. It can be applied to reproduction.

[0097]

As described above, according to the coded image transmission system of the present invention, the decompression process is performed by identifying the fixed parameters of the input image code, so that the inter-frame coded image in the input image code is processed. Since it is possible to specify the type, the processing for unnecessary inter-frame coded images that are not input or the motion compensation processing can be omitted, so that the decompression processing can be speeded up, and thus high-speed reproduction can be performed. Further, since the input image code itself is the same as that of the encoding system which is an international standard, an international standard image reproducing device can be used, and a dedicated image reproducing device can be eliminated.

Further, according to the image reproducing apparatus of the present invention, the process of determining the type of the inter-frame coded image and the process of decompressing the bidirectional inter-frame coded image in both the forward and backward directions can be omitted. And high-speed playback.

According to the third aspect of the present invention, the transmission interval of the intra-frame coded image is constant, and the type of the coded image can be determined based on the frame interval without using the header. You can save time.

According to the image reproducing apparatus of the fourth aspect, the processing time for motion compensation can be made unnecessary by performing the expansion processing of the input image code by omitting the motion compensation processing. Speed and high-speed playback.

Furthermore, according to the image reproducing apparatus of the present invention, the image of the previous frame is displayed as it is until the accumulated value exceeds the threshold value, and the time for displaying the inter-frame coded image at that time is displayed. Is omitted, the decompression process can be performed at high speed, and high-speed reproduction can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram of one embodiment of the present invention.

FIG. 2 is a diagram illustrating a hierarchical structure of an MPEG code format.

FIG. 3 is a configuration diagram of a code format conforming to MPEG.

FIG. 4 is an explanatory diagram of a picture fixing process.

FIG. 5 is a flowchart for explaining the operation of FIG. 1 based on a picture configuration.

FIG. 6 is a flowchart for explaining the operation of FIG. 1 when the picture interval is fixed.

FIG. 7 is an explanatory diagram of a motion compensation fixing process.

FIG. 8 is a flowchart for explaining the operation of FIG. 1 when there is motion compensation;

FIG. 9 is a flowchart illustrating the operation of FIG. 1 when there is no motion compensation.

FIG. 10 is an explanatory diagram of display update / non-update processing.

FIG. 11 is a flowchart illustrating an operation of updating a display according to an embodiment of the present invention.

FIG. 12 is a flowchart for explaining an operation at the time of updating / non-updating a display according to an embodiment of the present invention.

FIG. 13 is a block diagram of an example of a conventional device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 Header analysis means 12 Decoding means 13 Dequantization means 14 IDCT means 15, 22 Adder 16 Previous frame part 17 Rear frame part 18, 21 Pre-prediction means 19 Both prediction means 20 Post-prediction means

Claims (1)

(57) [Claims] An image is divided into a plurality of blocks and each block is divided into blocks.
Frames compressed and encoded using the discrete cosine transform
Intra-frame coded image and current frame for each block
And both the previous and next frames in time.
For motion compensation by searching for blocks with small differences
The current frame block and the motion-compensated block
Compression code using the discrete cosine transform
Multiplexes the encoded inter-coded image with the header
Replays the expanded image code and displays it on the display unit.
A decompression processing means for decompressing and reproducing an input image code in the raw device, and the inter-frame encoding decompressed by the decompression processing means.
Difference for calculating the cumulative value of the sum of image difference values for each frame
The sum total calculating means compares the cumulative value with the threshold value , and determines whether the cumulative value is equal to the threshold.
Output image data of the decompression processing means until the
Yes and a control means for stopping the display of the update on the display unit
An image reproducing apparatus, comprising: